Re: [PATCH v2] sched, timer: Use atomics for thread_group_cputimer to improve scalability

[Jason Low <jason.low2@hp.com>]

On Thu, 2015-03-05 at 16:35 +0100, Frederic Weisbecker wrote:
> On Mon, Mar 02, 2015 at 10:42:11AM -0800, Jason Low wrote:

> > +/* Sample thread_group_cputimer values in "cputimer", copy results to "times" */
> > +static inline void sample_group_cputimer(struct task_cputime *times,
> > +			                 struct thread_group_cputimer *cputimer)
> > +{
> > +        times->utime = atomic64_read(&cputimer->utime);
> > +        times->stime = atomic64_read(&cputimer->stime);
> > +        times->sum_exec_runtime = atomic64_read(&cputimer->sum_exec_runtime);
> 
> So, in the case we are calling that right after setting cputimer->running, I guess we are fine
> because we just updated cputimer with the freshest values.
> 
> But if we are reading this a while after, say several ticks further, there is a chance that
> we read stale values since we don't lock anymore.
> 
> I don't know if it matters or not, I guess it depends how stale it can be and how much precision
> we expect from posix cpu timers. It probably doesn't matter.
> 
> But just in case, atomic64_read_return(&cputimer->utime, 0) would make sure we get the freshest
> value because it performs a full barrier, at the cost of more overhead of course.

(Assuming that is atomic64_add_return :))

Yeah, there aren't any guarantees that we read the freshest value, but
since the lock isn't used to serialize subsequent accesses of
times->utime, ect..., the values can potentially become stale by the
time they get used anyway, even when we have the locking.

So I'm not sure if atomic64_add_return(&time, 0) for the reads would
really provide much of a benefit when we factor in the extra overhead.